In past years this report has concentrated first on how many people are going online for news and whether that has changed. Some of that is covered in the survey that PEJ and the Pew Internet and American Life Project produced together in early March 2010.
We also, however, wanted to dig more deeply into people’s online behavior: where they are going for news and what they do when they get there.
To do this, PEJ conducted an in-depth analysis using Nielsen’s NetView tool, a database of Nielsen’s online audience measurements.
That analysis shows that the websites of legacy news organizations – especially cable stations and newspapers – dominate the online news space in traffic in loyalty. But no one keeps visitors very long. The average visitor spends only 3 minutes 4 seconds per session on the typical news sites.
The metrics online are still evolving. Different services use different methods for measuring audience and behavior. Some in the online world argue that the methods by the two more traditional measuring companies, Nielsen and Comscore, tend to undercount traffic because of various methodological obstacles including access limitations to corporate servers during the day. Yet whatever the complications, most online professionals have told us that they believe the comparative relationships among different sites — the rankings and some of the patterns of usage that are uncovered in work like Nielsen’s — are useful and revealing, even if it is the case that overall traffic numbers are higher.
We began with Nielsen’s list of “News and Information” websites. We culled the list to remove sites that would not be considered news related, such as databases, consulting firms and sources like the National Weather Service. We also drove more deeply to identify individual news sites rather than families of sites. For instance, when Nielsen publishes its list of the top news sites, it uses CNN Digital Network, which is actually several different sites. In our analysis we use CNN.com, its primary news site. Next, we identified which of those sites averaged at least 500,000 unique visitors per month across three months in the fall of 2009 (September through November).
That analysis led to a range of findings:
- When it comes to online news, while it is a long tail world, with thousands of sites offering news, the top websites dominate traffic.1 Across the full roster Nielsen examines — the entire unfiltered list of 4,600 news and information websites — the top 7% collect 80% of the overall traffic.2 And the power curve even applies to top news sites. If we look just at news sites, refining the broad Nielsen list to eliminate sites from government and consulting firms that include information but not news, the same thing holds true. Among news sites that attract 500,000 monthly visitors or more, the top 10% of the most popular sites attract half the traffic.
- Legacy media still make up the majority of the most trafficked destinations (by Unique Audience),although newly created websites are joining the list each year. Of the top 199 sites in our analysis, 67% are from legacy media, and they account for 66% of the traffic. In all, 48% are from newspapers, and 19% from all other legacy media. Online-only sites make up the 33% balance of the list. The traffic figures closely match the raw numbers: legacy outlets get two-thirds of the traffic and online-only one-third (34%).
- Aggregators make up 27% of the top news websites and a few are among the most popular of all. Four of the top six news sites are either pure aggregator sites (Yahoo and Google News) or include a strong element of aggregation with some editing or original content (MSNBC Digital Network and AOL).
- The data also suggest that contrary to what some believe, specialty sites like those covering health care or science do not draw an especially loyal audience. Among the 39 niche sites on this list, all but two have audiences that are not only smaller, but also don’t stay as long, come back less often and don’t look at as many pages on the site as do audiences for either national and international sites or local sites.
- Two political sites, Daily Kos and Drudge Report, stand out with vastly higher numbers than the average site. Daily Kos averages 48 minutes per person per month and the Drudge Report nearly an hour, five times the average news site Given that minutes per person per month is partly a function of audience size, it is even clearer that these sites have a loyal audience. Neither site is in the top 30 in terms of audience, but in terms of time spent on the site per month they are No. 1 and No. 2.
- Among the biggest sites, the different missions of aggregators and originators of news are evident in the traffic patterns. Visitors to top news originators like the New York Times and CNN stay on average more than a minute longer per session than they do at aggregator sites like Google and Yahoo. They also tend to view more pages. But which type of site does a better job of getting people to return more often is less clear. And all the differences diminish as you move down the list to smaller sites.
To do this analysis, PEJ researchers examined these 199 websites in three different ways. First, we determined each website’s affiliation. Was it tied to a legacy outlet like print or television, or was it online only? Second, what was its editorial or topic focus? Was it national and international news, local or specialized around a certain topic and if so which one (science, celebrity, etc.)? Third, what was the nature of its content? Was it mostly reporting that the website produced itself? Was it mostly an aggregator of others’ work? Was it mostly offering commentary as opposed to news reporting?
For more info on how PEJ came up with this list, click here.
The top 20 sites at the front end of news sites are traditional brands most familiar to many. Except for one new aggregator of local news, these sites have all existed for more than a decade, and 15 (or 75%) are tied to legacy media.
Number of Sites
The majority – 16 of the top 20 – focus mainly on national and international news. Yahoo News isat the top of this group, which includes the sites of the three cable networks (MSNBC Digital Network, CNN.com and FoxNews.com) and the other two commercial networks (ABC and CBS – NBC is a part of MSNBC Digital Network) as well the British News site, BBC.com (the numbers here reflect only the U.S. audience).
Just four of the top 20 are sites focused on local news, three of which are actually networks of local sites with the same owner – CBS, Fox or NBC. The other locally focused site, Examiner.com, aggregates local newspaper content from all across the country and let’s the user choose which area to display.
Not a single site in the top 20 covers a niche or specialty topic area like politics or the environment.
When it comes to the content, all 20 sites except Yahoo News, Google News and Examiner.com produce some of their own content. A number of these originator sites, though, like MSNBC Digital Network, do more editing of Associated Press wires and other outside content than they do originating their own reports.
|Rank||Website||Unique Audience (000)||Rank||Website||Share|
|2||MSNBC Digital Network||35571||2||www.cnn.com||3.34|
|13||Daily News Online Edition||6889||13||www.washingtonpost.com||0.69|
|14||CBS Local Stations Group||6576||14||news.bbc.co.uk||0.67|
|16||NBC Local Media||5678||16||www.tvguide.com||0.63|
|18||Fox O&O TV Stations||5217||18||www.time.com||0.60|
If one compares data from Hitwise, another online metrics firm, with that of Nielsen there are some differences based on methodology.3 Hitwise tracks at the Internet service provider level, whereas Nielsen creates panels of users like Nielsen families. Some of the differences in their lists come from Nielsen grouping sites together more than Hitwise. There are some definitional differences of what is a news site. But if one looks at their lists of top sites, the first 20, there are few differences.
Yahoo News is on top of both lists. MSNBC and CNN come next in different orders, which may be explained by the fact that Hitwise is measuring only www.msnbc.com, while Nielsen is measuring MSNBC Digital Network, a larger group of MSNBC sites.
The starkest differences are the Drudge Report and AOL. The Drudge report is the 6th most visited site on the Hitwise list, and on Nielsen it is No. 33. AOL goes from the No. 10 site on Hitwise, to No. 4 on Nielsen. One possible reason for the difference is that Nielsen’s sample, according to some, undercounts daytime traffic since many workplaces will not allow Nielsen software on their servers. Hitwise data take more workplace traffic into account (see backgrounder on the different methods of Hitwise and Nielsen for a more detailed discussion of their methodologies).
On both lists, however, the traffic also drops off sharply after the top few websites. According to Nielsen’s data, just eight sites average more than 10 million monthly unique visitors (as measured from September through November, 2009 by Nielsen). The sites at the end of the top 20, CBS News and BBC News averaged half that. And the difference between the first site and the eighth is about 30 million.
This highly concentrated traffic at the top also reinforces the survey findings discussed in Online Audience that a majority of online news consumers graze on the Internet, but not very far, regularly visiting between two to five sites. And these data suggest that many people do not go very far down the list of the most popular destinations.
Looking at usage habits, the advantage of these top sites as a group goes beyond straight audience numbers.
- Users of the top 20 sites also returned more often and looked at more pages on the sites than others that come after the top 20. They averaged 5 visits a month per person and 18 Web page views, while the remaining 179 averaged only 3.3 visits per person and 14 Web pages.
- With all of those return visits, people devote more time over the course of a month to top websites than to those lower down – an average of 15 minutes per person among the top 20 sites versus 10 for the remaining 179.4
- One figure that was not much higher among the top sites compared with other sites was the amount of time visitors spend per session. They averaged 3 minutes and 4 seconds per visit, nearly identical to the 3 minutes and 6 seconds across all 199.
In other words, it seems to be more total audience and repeat visits rather than longer sessions that makes the difference in overall penetration.
The sway of legacy media remained strong, but less so, as one looks past the top 20 to the full list of the most popular news sites (those with at least 500,000 monthly unique visitors). Of these 199 sites, 33% were –online-only— (versus 25% of the top 20) while the other 66% had ties to a traditional platform.
Number of Sites
* Local TV includes networks of stations such as those owned and operated by Fox. These stations websites are combined into one figure.
In our longer list of 199 top news sites, cable news operations stood out both in total audience and in usage patterns. Each cable site averaged at least 23 million unique visitors monthly.
Newspapers, represented nearly 20 times more sites than cable, but commanded only about three times the combined audience, 181 million unique visitors.
And the traffic to the top three newspaper sites – the New York Times followed by the Washington Post and the USA Today – was half that of the three cable channels (37 million). It is also less than the top aggregator site alone, Yahoo News, which has close to 41 million.
Not only do cable sites get large audiences, but their audiences also return more frequently, go deeper and spend more time on the site per month
- Users spend more than twice as much time on cable sites than on newspaper sites or local TV sites. The cable sites average close to 24 minutes per person each month, versus only 10 minutes on both newspaper and local TV sites. This gap stands out even more as the newspaper sites in aggregate have three times the combined unique visitors as the cable sites. Thus, cable sites more than make up for any factor that total audience plays in the equation.
- Online only sites only did slightly better, just over 12 minutes per person each month, still half that of cable sites.
- The same phenomenon occurs when it comes to the pages viewed each month. Audiences to cable sites see twice the number of pages as those going to newspaper or local TV sites and nearly twice that of online sites.
|Sector||Monthly Visits per Person||Monthly Web Pages per Person||Monthly Time per Person (min:sec)|
- Looking at the magazine websites on this list, audience patterns run contrary to their association with more long-form journalism. These sites averaged just 6 minutes per person each month and less than 4 monthly page views.
Another way of categorizing the growing array of online news options is in the type of information they cover most. Which topic areas draw the greatest audience, the deepest website penetration or the most loyal audiences? To look for answers we broke the group into three main topic areas, according to the subjects they emphasize the most on the home page: 1) National and international affairs, 2) local news or 3) niche topic areas like health or politics, celebrity or science.
Over all, we found national and international sites got the most traffic. They also, more surprisingly, have a more loyal and engaged audience than specialty sites. Not only are the audiences to niche sites smaller, but users also don’t stay as long, don’t return as often and don’t go as deep into the site as do users of national and international sites or local ones.
- Local sites make up the biggest portion of our list of the top 199 sites: 87 sites (44%) but, combined, they account for only 25% of the traffic.
- National and international sites account for slightly fewer of the sites, 74 (37%) and more than double the audience – 65%. Their dominance in traffic suggests the degree to which assembling a broad range of topics, even if covered elsewhere as well, can also draw powerfully large traffic.
There were just 38 niche sites on the list of top sites and they drew just 10% of the audience.5
The other way to look at sites by topic was not just the size of the audience but how people used them. How often do they visit? How long do they stay? And many pages do they look at per visit?
Here, averaging the sites in each category, niche websites appeared to attract less committed traffic, contrary to what some might have expected. In general, people visited niches sites less often, stayed less time and did not view as many pages per visit as they did national and international or local news sites.
- Niche sites averaged half the time per person each month as did national and international news sites – 7 minutes versus 14 minutes.
- The same is true for how many pages users look at: people average 9 pages on a month on niche sites versus 17 pages on national and international sites.
- And people also return to niche sites less often over the course of a month– 2.9 times per month for niche versus 4.6 for national and international
Those findings raise some questions about the idea that people in the online world are simply digging deeper into a handful of subjects at the expense of getting general news. At least when it comes to the sites that generate the most traffic, these data suggest it is not that simple.
As Bill Densmore director of the Media Giraffe Project at the University of Massachusetts sees it, “the dominance of general-interest rather than niche sites suggests the enduring value of curation, or editing — that people like someone helping them make decisions about what’s important.”
|Content Emphasis||Monthly Visits Per Person||Monthly Web Pages Per Person||Monthly Time per Person (min:sec)|
|National and International||4.55||17.45||13:48|
Although these niche sites make up a small portion of the sites overall, we wanted to examine them a little more deeply to see which specific topics they covered and whether audiences use some niche sites differently than others.
Among the total 39 niche topic sites, 17 are about politics. Nine of these were conservative-leaning sites. Three were liberal, and five were non-partisan.
A full analysis of blogs might reveal something different, but many either are not considered news and information sites by Nielsen or fall short of the minimum audience level (500,000 unique visitors a month) to be included in this data set. Even some that the tracking firm Technorati ranks as the most influential – which we analyze in a separate section of this report, – fall short of the minimum audience level.
But among the top 199 sites here there, are 22 that could be considered blogs, including Michele Malkin.com, Breibart.com and Daily Kos. While a separate analysis of blogs would be worthwhile, this list at least suggests the idea that the Internet moves left while things like talk radio tilt right, isn’t quite the case. Of these 22 blogs, three were liberal and eight were conservative. The rest were non-ideological or did not deal with politics.
Political sites tended to have higher user engagement than news sites over all, while sites that focus on other topics like entertainment, science and technology — fall below the average for most usage measurements.
- The average user seems far more engaged in politically leaning sites, compared to all news sites. Users spend about 23 minutes per person per month on the liberal sites and 20 minutes on conservative sites. This compares to only 12 minutes on the average site.
- Users also view more pages on political sites than among news sites over all. The average user looks at 25 pages on liberal sites and 30 on conservative ones.
This same pattern continues with how many times users return to the sites and how much time they spent per visit to the site.
|Monthly Visits per Person||Monthly Web Pages Per Person||Monthly Time per Person (min:sec)|
Part of these high averages is due to two sites, the liberal blog site Daily Kos and the Drudge Report, a news aggregator that offers a smattering of occasional reporting with a distinctly conservative flavor. Daily Kos averages 48 minutes per person per visit and the Drudge Report nearly an hour, five times the average news site. These two sites sharply increase the overall numbers for politically leaning sites. The other sites range between 5 and 25 minutes per person, still above the average but to a much lesser degree.
The two other most popular niche topic areas have much less user engagement, compared to both political sites and news sites overall.
- Science and technology sites averaged just under four minutes per person each month, much lower than the average news site (12 minutes). They also averaged fewer return visits per month, just 1.97 per person in a month.
- Entertainment sites faired slightly better with an average of just under six minutes per person and 2.6 visits per person per month.
- Users are also going looking at fewer pages on these sites than the average news site, with 9 Web pages per person per month for entertainment sites and 5 for science and technology sites, compared to 16 for the average site.
|Monthly Visits per Person||Monthly Web Pages Per Person||Monthly Time per Person (min:sec)|
|Science and Technology||1.97||5.50||03:54|
PEJ also categorized the top 199 sites by how their content is organized and presented: originator, aggregator, or commentary. In making these categories PEJ looked at the front page of each site and counted the links on the site. If two-thirds of the links on the site were original content, the site was labeled an originator. If two-thirds of the links were to outside content (including the Associated Press and other wire services), the site was categorized as an aggregator. Commentary sites are those that do not have original content in terms of original reporting, but have content that is mostly commenting or discussing reporting done by others.
AOL, for example, is categorized as a content originator although it began as largely an aggregator. In the analysis conducted in January, two-thirds of the links on its front page were to content produced by AOL staffers. But through the course of 2009, before it had added staff to produce original content, PEJ’s content analysis found that 84% of its top stories were actually copy produced by others.
Others on this list fall into the originator category, but might strike many as a hybrid of original work and aggregation. MSNBC Digital Network is an example. Most of the stories on its site contain bylines or credit lines attributed to MSNBC. But most of the stories at the bottom also note that there is wire copy that also contributed to each of these accounts. Since its redesign, most copy on CNN’s website carries byline from CNN staff. In the past the site relied heavily on wire copy for its most up-to-date news. Rules at the Associated Press allow news organizations to use their own byline even if only minimal alterations were made to the AP copy.We cannot assess to what extent a piece of content might be derived from wire material. This study, in other words, took a site at its word when it claimed credit for original work.
A site like Huffingtonpost, which is a commentary site, does do original reporting; the bulk of the content on its home page, however, is commentary, most of which links to original reporting done by others.
Of the 199 news sites, 27 (14%) primarily aggregate information produced by others; 20 (10%) put most of their resources toward offering commentary about news events first reported on by others. The largest portion, 152 (76%) produces their own original content.
The aggregators included here range from algorithmic sites like Bing News to Buzzle.com which is aggregated, by an individual.6
The aggregator and commentary groups are, predictably, dominated by online-only sites: 19 of the 27 aggregator sites and 19 of the 20 commentary sites.
Among those that originate content, the majority, 124, are affiliated with legacy media outlets. Of these sites the largest sub-group is newspapers, with 90 sites, followed by local TV-based sites, 10. Many of the local TV sites, though, represent groups of stations, such as CBS owned and operated stations, which could not be disaggregated. Thus their audience figures are not directly comparable with a single site like the CNN.com. The remaining 28 originated as online only entities.
When it comes to usage patterns though, the different missions of aggregators and originators of news are only evident to a certain degree. Among the very top sites, some distinctions emerge, but others seem tied little to the type of content being offered.
Among the top ten sites studied here, visitors to news originators average more than a minute longer per session (visit) that at aggregator sites like Google and Yahoo. They also tend to view more pages. But what is less clear is which type of site does a better job of getting people to return more often. Here, CNN and Fox stand out above all the rest, but Yahoo scores higher than most other news sites.
|Website||Unique Audience (000)||Sessions per Person||Monthly Minutes per Person (min:sec)||Monthly Minutes per Session (min:sec)||Page Views per Session||Monthly Web Pages per Person|
|MSNBC Digital Network||35571||5.66||19:18||3:25||4.21||24|
Also, as mentioned above, among the news originators at the very top, the three cable sites stand out from the newspaper sites, commanding roughly twice the time spent and visits made per person.
When looking at aggregators and news producers across the entire group of 199 sites, the differences diminish. Here, the minutes per person vary only slightly as do the depth of pages viewed. The only behavior measurement that shows much variation at all is the average monthly visits per person. But this can be as much a function of the total audience size as individual behavior patterns.
Commentary sites faired the worst with users spending around 11 minutes on commentary sites in November, compared to 13 on an aggregator or originator.
|Content Type||Monthly Visits per Person||Web Pages per Person||Monthly Time per Person (min:sec)|